HAZARDS OF LINEARIZATION OF LANGMUIR MODEL

被引:18
作者
HARRISON, F [1 ]
KATTI, SK [1 ]
机构
[1] UNIV MISSOURI,COLUMBIA,MO 65211
关键词
D O I
10.1016/0169-7439(90)80075-H
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
We study Langmuir's isotherm. The response variable is Y, and the control variable is x. The isotherm is Y = ax/(1 + bx) + eplison where epsilon is independently and identically normally distributed, N(O, sigma-2). We studied linearized models, such as the Scatchard plot, of Y/x versus Y, and the double reciprocal linearization. The Scatchard plot had a 90% midrange which did not contain the true curve in a vast portion of the range. The double reciprocal linerization has undefined expected values and gets worse as the sample size increases. We give a method to obtain direct nonlinear least-squares estimates. Based on a study of 32000 simulated samples, we recommend regression without transformation of the model.
引用
收藏
页码:249 / 255
页数:7
相关论文
共 6 条
[1]  
BAILEY JE, 1986, BIOCH ENG FUNDAMENTA, P106
[2]  
COLQUHOUN D, 1971, LECTURES BIOSTATISTI, P257
[3]  
HABIBULLAH M, 1988, THESIS U MISSOURI CO
[4]  
LEHNINGER AL, 1985, BIOCHEMISTRY-US, P191
[5]  
OLIVER FR, 1970, APPL STATIST, V19, P269
[6]  
SNYDER LR, 1968, PRINCIPLES ADSORPTIO, P58